Emptiness and Positionlessness
Do the Madhyamika Relinquish All Views?
IN THE FINAL VERSE of his major work, Mulamadhyamakakarika, Nagarjuna writes: I prostrate to Gautama
Who through compassion
Taught the true doctrine,
Which leads to the relinquishing of all views. (27: 30)
This verse echoes an earlier remark:
The victorious ones have said
That emptiness is the relinquishing of all views.
For whomever emptiness is a view
That one has accomplished nothing. (13: 8)
It is not at all clear how one is to understand such claims. One response, of course, would be to mutter either with reverence or with derision about the mystery, paradox, and irrationality or about the fantastic transcendence of rationality of the East. But either would be too facile. For one thing, it would be simply to shirk the hermeneutic burden of making sense of the text. But more compellingly, given the relentlessly rational, meticulously argued character of the text as a whole, and indeed of his entire corpus, such a reaction would be hard to justify with respect to Nagarjuna.
And finally, given the extensive commentarial tradition rooted in this text, beginning with Nagarjuna's immediate disciple Aryadeva and continuing to the present day, with its different readings and interpretations, several competing understandings of the text are already available. So we can't avoid the task of providing a coherent reading of these prima facie paradoxical verses, nor can we ignore their apparently paradoxical character. For Nagarjuna is, throughout his philosophical corpus, in one straightforward sense, advancing a view, the Madhyamaka view that all phenomena are ultimately empty though conventionally real and dependency originated, and so on.
And it is through adopting this view that Nagarjuna holds that we can attain nirvana— liberation from the suffering whose root lies in false views. Can Nagarjuna be seriously suggesting that we relinquish that view? Would that not be selfdefeating? Would that not be like saying out loud, "I am now silent"? A long and highly influential line of commentators, beginning with rje Tsongkhapa, the founder of the dGe lugs school of Tibetan Buddhism, and his disciple mKhas grub rje, provide a straightforward solution to the hermeneutic problem these verses pose: they simply argue that when Nagarjuna speaks of relinquishing "all views," he means "all false views," or "all views according to which things are inherently existent."
Nowadays some who wish to be prasarigika-madhyamikas say: Our system even conventionally does not have any assertions based on the ultimate or the conventional.... Therefore there is no such thing as an own system for prasangikas since Nagarjuna, Aryadeva, and Candraklrti say that madhyamikas have no position and no thesis. As Nagarjuna's refutation of Objections says:
29 If I had any proposition (pratijna) then this defect would be mine. I have, however, no proposition. Therefore there is no defect that is mine.
Answer: If this which you propound is not the Madhyamaka system, then it is contradictory to establish it through citing passages from the superior Nagarjuna and his spiritual sons.... If you say it is Madhyamaka and, from within that, the system of Candraklrti, then it would contradict your assertion that madhyamikas in general and Candraklrti in particular do not have their own system ... (LRCM 435 b in Hopkins 1983) Since it would have to be that there was no correct view leading to the state of nirvana, all the activities of hearing, thinking etc.... with respect to the Madhyamaka scriptures would be senseless. (TKP: LRCM 471 b$ in Napper 1993)
The text, they point out, is full of passages in which the assertions are subject to implicit qualification of this kind. So, for instance, when Nagarjuna in chapter 2 says, So movement and motion
And agent of motion are non-existent. (2: 25)
nobody thinks that he is denying that anything ever moves. Rather, all commentators agree that he is asserting that we can't think of motion, movers, or the act of moving as inherently existent or independent. The dGe lugs pa commentators emphasize that to do this, we read the word nonexistent as qualified implicitly by inherently—that is, as "not inherently existent."
So, given the fact that Madhyamaka hermeneutics regularly requires the interpretive insertion of such qualifiers, and given that the passages can be readily demystified with just such an insertion, there is strong reason to read them in this way.
But there is more. Emptiness, for a madhyamika, is an ultimate truth. One can achieve a correct view—a view of things as they in fact are. Such a view surely should not be relinquished, for this would be to relinquish the soteriological goal of all of Buddhist practice. So, the dGe lugs pas argue, one must read Nagarjuna as suggesting straightforwardly, rationally, and without even a hint of paradox, that one should relinquish all false views, and that for the one who views emptiness as inherently existent there is no hope.
Though this route has every hermeneutic virtue, it emerges in the fourteenth century as a response to an older and then prevalent view, expounded most forcefully in Tibet by Ngog bio Idan shes rab, whose own texts unfortunately do not survive except in fragments quoted in later works. This older view is dominant in the commentarial tradition of the Nying-ma school of Tibetan Buddhism. According to Ngog, Nagarjuna means just what he says. The central teaching of Madhyamaka is that one should relinquish all views, and that if Madhyamaka becomes a philosophical view, one has fundamentally missed its point.2
Scriptural evidence for Ngog's view includes this passage from Santideva's Bodhicaryavatara: 2 The conventional and the ultimate
Are explained to be the two truths.
The ultimate is not grasped as an object of thought;
Thought is explained to be merely conventional.
(Santideva, Bodhicaryavatara 10)
This of course requires the correlative and equally challenging theses that one who knows emptiness has no view of anything—and hence has objectless knowledge, and that emptiness itself is not an object of knowledge, and not an entity. And Ngog and his Nying-ma followers do not shy away from these claims.3
Making sense of these claims is not easy, but this is my intention in this chapter. For Nagarjuna generally, the dGe lugs to the contrary notwithstanding, says what he means. And, in the verses in question, nothing forces the kinds of implicit qualifiers that often are contextually forced. Too much of the remainder of the text, as well as Nagarjuna's reply in his Vigrahavyavartam (to be discussed later in this chapter), coheres with the straightforward literal reading of these verses.
Moreover, reading Nagarjuna in this way highlights some of the intriguing parallels between Madhyamaka philosophical method and Western Pyrrhonism, as it is articulated in the tradition running from Pyrrho and Timon through Sextus Empiricus right up to Wittgenstein's Tractatus and Philosophical Investigations. While subjecting an Indian Buddhist text to a European procrustean bed
is hardly by itself an argument for an interpretation, there are so many powerful affinities in method and in detail between Madhyamaka and Western skepticism that the appearance of a further parallel must be taken as at least suggestive. So when we consider remarks like the following, we are forced to ask ourselves how we in the West make coherent sense of them, and whether the same interpretive strategies might not be apposite in the case of Nagarjuna: We determine nothing.
All things are undetermined.
We must not say about any one thing that it is or that it is not or that it is and it is not or that it neither is nor is not. (Pyrrho)4 [T]he sceptic does not take the real existence of these formulae wholly for granted. As he understands them, the formula "All things are false," for example, asserts its own falsity together with that of other things.... (36) [I]f dogma is denned as "assent to a non-evident thing," then we shall say that we have no system. But if one means by "system" a "discipline which, in accordance with appearance, follows a certain line of reasoning, indicating how it is possible to live rightly... we shall say that we do have a system." (37)
Non-assertion ... is a disuse of assertion in which ... both affirmation and denial are applied.... [W]e neither affirm nor deny anything. (81) (Sextus, in Hallie 1985)
6.54 My propositions serve as elucidations in the following way: anyone who understands me eventually recognizes them as nonsensical, when he has used them—as steps—to climb beyond them. (He must, so to speak, throw away the ladder after he has climbed up it.)
7 What we cannot speak about we must pass over in silence. Wittgenstein, (Tractatus, trans. B. F. McGuiness, 1922)
What these traditions share, at the broadest level of description, is the philosophical project of undermining essentialism. For Pyrrho and for Sextus, Wittgenstein, and Nagarjuna, and for all of their skeptical followers, the fundamental philosophical error is to propose a characterization of the nature of things. This is so, from their perspective, not because the nature of things is elusive, but rather because there is no nature of things—because the very concept of an essence is itself incoherent.
In reading Western skepticism, we must always be sensitive to the expository difficulties of this project: for essentialism is virtually built into the grammar of our language. That is why it is so seductive. To articulate the critique requires a careful account of how language works, and of how it is being used in the philosophical critique in question.
And that account itself will be subject to the same misconstruals. But we can understand what it is to kick away the ladder, and we can question the primacy of assertion as a linguistic act. This kicking and questioning—this banging our heads against the walls of language—is essential to attaining clarity about the role of language and conception in ontology and in our mode of being in the world (see chapters i and 5 of this volume for more on this issue).
This questioning of the primacy of assertion and with it the realist semantics for natural language that undergirds this presupposition lies at the heart of the Madhyamaka critique of essentialist philosophy. Such probing also affirms the important affinities of this ancient Indian philosophical movement to much of the contemporary, postmodern philosophy of language and epistemology based on the critical work of Wittgenstein.
The affinities extend, as one would expect, beyond the narrow confines of semantics, to indicate a general agreement regarding the inescapability of situatedness and perspective in knowledge.
They also share a faith in the possibility of a recognition of that predicament and in the power of that recognition to liberate us from the tyranny of our particular perspective. My aim in this chapter is not only to defend one side of a recherche medieval Tibetan debate in Nagarjuna interpretation, but also to illuminate these larger issues in the philosophy of language and epistemology. So let us work towards such an understanding of Nagarjuna's claim to relinquish all views, with Pyrrhonism in the corner of our eye, and more of Nagarjuna's enterprise in view. We begin with a survey of emptiness itself.
EMPTINESS
When Nagarjuna and his followers assert that all phenomena are empty, it is important to ask just what phenomena are asserted to be empty of. For emptiness is not, of course, a monadic property. A room that is empty of elephants, such as the one in which I am now writing, may not be empty of people. The property of which Nagarjuna asserts phenomena to be empty is most often referred to as "inherent existence," which denotes the putative property of existing simply per se, independent of any conditions or relations or other phenomena.
But we can also say that it is to be empty of essence—empty of any intrinsic property or characteristic that makes an entity what it is, or to be empty of substantial existence—to be empty of being an entity distinct from and a basis of its attributes. A good deal of Madhyamaka philosophy is devoted to showing that these characterizations amount to the same things.
The assertion that all phenomena are empty is tightly bound up with the doctrine of the Two Truths—of a conventional truth and an ultimate truth, and of the complex relation between them. The ultimate truth about phenomena is their emptiness: their lack of essence, independence, and exist ultimately would be to exist substantiality substantially independently, by virtue of the possession Emptiness and Positionlessness
of an essence. Nothing, the madhyamika asserts, exists in this way. But that does not mean that all phenomena are completely nonexistent, that they are imaginary. Rather, argue Nagarjuna and his followers, real phenomena are conventionally existent. To be conventionally existent is to exist dependently, to possess one's identity nominally, to be essenceless and impermanent. A thorough discussion of the doctrine of the emptiness of phenomena and of the connection between ultimate emptiness and conventional or nominal reality would take us far afield. For now, these remarks will have to suffice.5
It is crucial to realize that Nagarjuna does not argue that all conventional phenomena are empty of essence only to posit emptiness itself as substantially existent. Instead emptiness itself, rather than existing as an absolute reality beyond the relative—as true existence as opposed to illusion—is no more than the emptiness of conventional phenomena. As such, it too, is dependently originated, merely conventionally existent, and hence empty of inherent existence. This doctrine of the emptiness of emptiness, and consequent refusal to ascribe inherent existence to anything—whether a conventional or an ultimate phenomenon—is common to all Madhyamaka schools.
Emptiness must hence be understood, in this philosophical context, neither as nonexistence nor as an independent reality lying behind a veil of illusion, but rather as the mode of existence of conventional phenomena, and at that a mode that exists in exactly the same way as any other conventional phenomenon.
Emptiness is, in short, nothing more than the fact that conventional dependent phenomena are conventional and dependent. It is simply the only way in which anything can exist. Rather than to assert that because phenomena are empty they are nonexistent, the madhyamika asserts that because phenomena exist they are empty.
For this reason it is important not to move from the fact that all phenomena are empty of essence to the assertion that emptiness is the essence of all phenomena. That is, we do not want to turn essencelessness itself into an essence. For if it is correct to argue that all things are empty of essence, and if emptiness were an essence, it would paradoxically turn out that all things were nonempty, in which case they would have essences.
This purely negative characterization of emptiness must be kept in mind as we examine the degree to which and sense in which emptiness can be an object of knowledge or an entity, and so the sense in which Madhyamaka philosophy must be understood as, in the end, positionless.7
(See chapter 5 of this volume for an extended discussion of the paradoxes that lie in this neighbourhood.)
EMPTINESS AND MEANING
Nagarjuna really advocates relinquishing all views is in fact semantic: is it possible to understand the words that Nagarjuna utters to claim that he is not expressing propositions? If not, either he is flatly self-contradictory, talking nonsense, or in fact means something else. If so, I will have identified a nonassertorial understanding of his words, according to which they are nonetheless true. We have seen that, from the standpoint of Madhyamaka, there is no convention-independent reality correspondence with which could be truthmaking.
But that does not entail the impossibility of a correspondence theory of truth tout court. Indeed, the very emptiness of phenomena that makes correspondence with a convention-independent reality chimerical ensures the empirical existence of a world of conventional entities. And indeed, these entities are our referents when we use ordinary, or even philosophical, discourse. They are the domain of conventional truth. For Nagarjuna and his followers, sentences are indeed conventionally true just in case the entities their referring terms designate in fact satisfy their predicates, where these terms are understood in the ordinary way.
But what about sentences that purport to characterize ultimates or to describe things from the standpoint of ultimate knowledge? Here things are a bit more delicate. The subject, now, remember, is sentences that make assertions about emptiness itself, sentences like "Emptiness is itself empty," or perhaps more interestingly, "Emptiness is without characteristics."
In thinking about these sentences, one must also bear in mind that their status must be considered from two distinct vantage points. The first is that of the ordinary philosopher whose objects of direct apprehension are conventional entities, but who can inferentially realize their emptiness, and who, by extension, can be the subject of cognitive states such as beliefs whose direct object, via conception, is emptiness.
The second is that of the arhat, who directly perceives emptiness as it is unmediated either by apprehension of (other) conventional entities or by conception. First things first: given the Madhyamaka analysis of the emptiness of all phenomena, and especially given the particular working out within Madhyamaka philosophy of the identity of the Two Truths and of the emptiness of emptiness, even ultimates are empty of inherent existence, and hence dependently originated, and hence merely conventional.
So when I say that emptiness is empty of inherent existence, if you were to ask about the ontological basis of the truth of that sentence, at one level, the story I can tell is just the same as the one that I tell regarding the ontological basis of the truth of my claim that snow is white. Emptiness, like snow, is ultimately empty but (and therefore) conventionally existent.
Insofar as it is conventionally existent, we can ask about the nominal properties that conventional entities can be asserted to have, and among them is, in the case of emptiness, its emptiness, just as among the properties that we can say snow has is its whiteness. But in neither case do we suppose, in order to make sense of this ordinary discourse, either the substantial existence of the referent of the Emptiness and Positionlessness
subject term, or the independent existence of the ascribed property, or anything besides the nominal truth of the sentence. That is what it means to hypostasize neither the substantiality of conventional entities nor of the lack thereof. But things, as we shall see, are subtly different from the standpoint of one directly perceiving ultimates.
To make this case completely and to spell out its contents in detail will take some time and require a few side trips through semantic and epistemological terrain. But to explain my motivation for these detours, I offer a preliminary sketch of the destination (a sort of soteriological postcard: having a great time; wish you were here!): for the arhat who directly realizes emptiness, nothing is present in consciousness but emptiness itself. For such a consciousness there literally is no object, since there is in such a consciousness no reification of the kind that gives rise to subject-object duality.
Moreover, since such a consciousness is directed only toward what can be found ultimately to exist, and since nothing can be so found, there is literally nothing toward which such a consciousness can be directed. But this very fact is what is ostended by the dictum that emptiness is itself empty: That is the fact that emptiness is not the real object as opposed to the unreal object of ordinary perception; not the object that appears when false appearance is shed; that it, to the extent that it appears as an object at all, it does so as falsely as does any table. But then the semantic story told from this perspective is somewhat different.
The best that can then be said is that from such a standpoint the words "Emptiness is empty" ascribe no property to any object at all: from that standpoint, they express no proposition. At that stage, there is no view to be expressed, where a view is something that can be given assertoric voice. Or so I will argue. This is the point that I take Nagarjuna to be seeking when in Vigrahavyavartam he writes:
29 If I had any proposition at all,
Thereby I would have that fault.
Since I don't have a proposition
I don't have any fault at all.
If I had even one proposition thereby it would be just as you have said. Though if I had a proposition with the characteristic that you described, I would have that fault, I have no proposition at all. Thus, since all phenomena are empty, at peace, by nature isolated, how could there be a proposition? How can there be a characteristic of a proposition? And how can there be a fault arising from the characteristic of a proposition? Thus, the statement, "through the characteristic of your proposition you come to acquire the fault" is not true.
And so let us now take some time to examine this locus classicus for textual support for the position I now defend. There is a persuasive strain of dGe lugs Before moving more directly to the question that concerns us, I turn to a few points regarding Madhyamaka philosophy of language. For the question whether pa and dGe lugs-inspired Western scholarship that urges the insertion of some qualifiers here—that Nagarjuna really wishes to renounce only inherently existent propositions, or propositions that ascribe inherent existence to their objects. This interpretation derives some support from one contextual dimension. Consider, the dGe lugs say, the opponent's charge in the verse to which verse 29 responds: 4 If you say it is like the negation of a negation,
If you maintain that, that won't do any good.
Seeing it that way, your proposition will thus
Have a defective characteristic, not mine.
If we pay attention merely to this dialectical move, the dGe lugs' construal is quite natural. The substantialist opponent charges Nagarjuna with using an assertion (that all phenomena are empty) that he must, in order to be consistent, regard as itself empty. Now, for the opponent, emptiness is nonexistence. So the opponent charges Nagarjuna with the self-refuting assertion that everything is nonexistent, including that very assertion. And she points out that the same
charge does not apply to anyone asserting the negation of the proposition—that all things are non-empty—that is, that phenomena in fact exist in some way, since for one making that assertion, the very assertion counts as evidence of its truth.
So on this reading, Nagarjuna is only replying to a criticism that misconstrues the nature of emptiness and asserting that on a proper construal the alleged logical error is not committed. But consider the auto-commentary on the opponent's move:
It may appear to you that a negation of a negation is impossible in this way. If so, your statement negating the negation of the statement that all things have essence would thus also be impossible. To this we reply: That won't do any good! This follows because the objection applies to you in virtue of the character of your proposition; since mine doesn't have that character, to say that the negation of a negation is not possible doesn't make any sense.
This subtle semantic remark requires close attention. It is important to bear in mind the Nyaya-influenced logico-semantic context in which these debates originate. The dominant view of the nature of meaningful assertion (the one that Nagarjuna questions) is one that from our perspective can best be characterized as a version of Fregean realism: Meaningful assertions are meaningful because they denote or express independently existent propositions.9
A proposition is the pervasion of an individual entity or groups of entities by a real universal or sequence of universals.
Against this background, opponent can be seen as attacking Nagarjuna at a deeper, more ontologically significant level: she anticipates that Nagarjuna will Emptiness and Positionlessness charge her with what must be, on her own terms, a contradiction.
For the opponent has, one move earlier in the debate, argued that Nagarjuna's words, if true, deny the existence of what could make them true, that is, languageindependent entities and universals. Hence, she asserts, Nagarjuna's words are nonsense.
But this, the opponent recognizes, presents her with a prima facie problem, by virtue of her presupposition of the existence of propositions as foundations of meaning: in leveling this charge of self-undermining against Nagarjuna, she in fact asserts the negation of Nagarjuna's thesis. But the negation of nonsense must itself be nonsense. Hence Nagarjuna, in this sketch of her next move in the debate has her anticipating his objection that the negation of nonsense must, given the opponent's position, then itself be nonsense, with the consequence that the opponent's position would be self-refuting.12
The opponent then replies that all she is doing is providing a reductio on Nagarjuna's position, and not offering an independent thesis of her own. In the context of Prasahgika-Madhyamaka methodology, the irony is palpable and the logic poignant.13 Now consider the auto-commentary on Nagarjuna's reply:
If I had even one proposition thereby it would be just as you have said. Though if I had a proposition with the characteristic that you described, I would have that fault, I have no proposition at all. Thus, since all phenomena are empty, at peace, by nature isolated, how could there be a proposition? How can there be a characteristic of a proposition? And how can there be a fault arising from the characteristic of a proposition? Thus, the statement, "through the characteristic of your proposition you come to acquire the fault" is not true.
Note that Nagarjuna does not refer here to emptiness or to nonexistence, or to any supposed confusion between the two. He is worried about whether or not he is asserting a proposition,14 and so he should be. For the opponent has tried to
turn the ontological tables on him, presenting a reductio whose conclusion is not that Nagarjuna's assertion that all things, including those very words, are nonexistent entails its own nonexistence, but rather that his words, if true, are meaningless, since if they are true no proposition corresponds to them. If Nagarjuna shared with her the presupposition that meaning is to be analyzed in terms of relations of words to propositions, the reductio would hit the mark. So he must respond by denying the commitment to such an analysis.
And that, of course, is just what he does: it is characteristic of Madhyamaka rhetorical strategy, and particular trademark of Nagarjuna, to refute an opponent's position by demonstrating that the uncomfortable consequence she attributes to the madhyamika's position not only does not attach thereto but moreover is a presupposition of the opponent's position. Here Nagarjuna, adopting that strategy, asserts that he is merely using words. Their emptiness amounts to their being merely conventional tools denoting nothing, and whose sense presupposes no language-independent denotation but rather derives merely from their utility in discourse.
In exactly this sense he has no proposition, though he uses plenty of words. It is the opponent, who, by virtue of denying the emptiness of language, is committed to the existence of propositions as entities that ground meaning. But if she is so committed, then she has a problem, for she must then make sense of her own denial of Nagarjuna's proposition. And that requires that she concede the coherence of Nagarjuna's thesis. But to concede its coherence is to give up the reductio. Moreover, it is to concede that meaning does not presuppose the existence of propositions, but rather relocates the burden of proof squarely on the opponent to demonstrate the existence of an unobserved metaphysical fifth wheel, and one that even leads to paradox.
This reading of these crucial verses from Vigrahavyavartanl establishes several important points that are essential for understanding the madhyamika's claims to relinquish all views:
(i) Nagarjuna does sincerely claim to assert no proposition, not merely to assert no inherently existent proposition.
(2) He does not thereby deny that he uses words.
(3) The claim to assert no proposition is a semantic claim that is bound up with the claim that language, like all other phenomena, is empty.
(4) Nagarjuna does not disclaim the truth of his utterances, though he does disclaim a nonconventional ontological ground for that truth.
EMPTINESS AS KNOWABLE I: CAN EMPTINESS BE AN OBJECT OF KNOWLEDGE?
At this point, I return to epistemological terrain and ask about the nature of our knowledge about emptiness. For positionlessness must be characterized in such a way that it does not, self-refutingly, entail that we cannot have the knowledge that gives rise to it. Moreover, of course, it must come out that it is consistent with all of the rest of Nagarjuna's assertions. So, to put the point most straightforwardly, it must come out true that things are empty, and this truth must be knowable. It will therefore be a constraint on our analysis that a madhyamika philosopher can say these things and come to realize the truth of emptiness in such a way as to enable release from the delusions of samsara.l5J6 Nonetheless, despite this emphasis on the knowability of emptiness, no party to this dispute would argue that emptiness exists inherently, or ultimately. Emptiness, all agree, is a conventional existent. But, as we will see in the discussion that follows, what it is to be a conventional entity is not a straightforward matter.
POSITIVE AND NEGATIVE TETRALEMMAS
I now add one more piece to this philosophical jigsaw puzzle. Nagarjuna, famously, uses the classical Buddhist catuhskoti, or tetralemma, form of analysis,
which considers with respect to any pair of contradictories not only the possibilities of each being true, but also those of both being true and neither. Interestingly, though, in Mulamadhyamakakarika, the tetralemma is deployed in two distinct ways, which we can call "positive tetralemmas" and "negative tetralemmas." When the phenomena under discussion are conventional entities, Nagarjuna deploys positive tetralemmas. Consider this example, for instance: Everything is real and is not real,
Both real and not real,
Neither real nor not real.
This is Lord Buddha's teaching. (MK 18: 8)
This tetralemmic verse occurs in the context of the discussion of the nature of the self and of external entities, a discussion that emphasizes the conventional reality of phenomena on both sides of that divide, together with their ultimate nonexistence. Here, Nagarjuna notes that each of the four branches of the tetralemma with regard to reality can be asserted, subject to appropriate ontological qualification:
(i) Everything is conventionally real.
(2) Everything is ultimately not real.
(3) Everything is both conventionally real and ultimately not real.
(4) Everything is neither ultimately real nor completely unreal.
The important point here is not simply that these four assertions are all held to be consistent but rather that it is appropriate to make these assertions in the first place. As we shall see in our next examples, that is not always, according to Nagarjuna, the case. "Empty" should not be asserted.
"Non-empty" should not be asserted.
Neither both nor neither should be asserted.
They are only used nominally. (MK 22: 11)
Having passed into nirvana, the Victorious Conqueror
Neither found existence evident
Nor found non-existence.
Nor both nor neither thus to be evident. (25: 17)
In the first of these two verses, Nagarjuna is discussing the status of emptiness in the context of a discussion of the nature of the Buddha and of Buddhahood. In the second, he is explicitly discussing the problem of formulating ontology from the standpoint of nirvana. In these cases, where the subjects regarding which we are tempted to make assertions are considered from the ultimate point of view—that is, from the epistemic standpoint of one who sees things as they are independently of conventions—rather than assert that all branches of the tetralemma can be asserted, Nagarjuna explicitly precludes assertion of any of them.
Each would be fundamentally misleading. We see here that Nagarjuna is drawing a logical distinction between two epistemological standpoints: as long as we remain within the conventional standpoint, we can, providing that we are careful, say many things, mundane and philosophical. But once we transcend that standpoint, no matter what we try to say, and no matter now carefully we hew to a via negativa, we can say nothing at all consistent with the via media Nagarjuna is determined to limn. This will provide a valuable clue to the sense in which Madhyamaka philosophy requires us to regard emptiness not as an entity, and to relinquish all views when we understand emptiness.
VIEWS AND ENTITIES
It is now time to make explicit the link between these two points—that regarding the status of emptiness as an entity, and that regarding the possibility of holding a view once one has directly cognized emptiness. The metaphor of view must first be explicated and unpacked a bit. Note that the ocular metaphor represented by the English term view is mirrored in both the Tibetan and the Sanskrit terms it is used to translate. In Tibetan, the term is Ita ba, literally "seen." In Sanskrit there are two terms, both translated by the same Tibetan term, and their difference will be important for us, since, as Ruegg (1977) notes, they differ systematically with respect to the issue at hand. Both are derived from the root drs, "to see." The first, the one used in the admonitions against holding views, is drsti.
The second, used when the cognitive relation to emptines that is the goal of madhyamika analysis and practice is being characterized is darsana, best literally rendered as "direct awareness", or "coming face to face with." Neither Nagarjuna nor his madhyamika followers ever deny the value or possibility of sunyata- darsana, (view of emptiness) though they are critical of the very idea of sunyatadrsti. So here, using sunyata- darsana" Candrakirti writes, The thorough extinguishment of attachment is the cause of attaining nirvana, and, except for the view of the lack of inherent existence, there is no other doctrine which is a cause of thoroughly extinguishing such attachment. (Prasannapada ii6a) Tibetan, as Ruegg notes, does not draw this terminological distinction. Nor does English.
The metaphorical content here is doing real philosophical work, even if covertly. For (bracketing for now the interpretation of sunyata- darsana) when we view something, there must be something that we view, and we must view it from some perspective, with awareness of it under some description.17
And so a view is possible if there is something to view and some way in which it is viewed. Let us take the first entailment, the ontic, first: if it were possible to have a (true) view about emptiness, emptiness would have to be a thing, an object of awareness. But if we supposed that it is, a dilemma emerges: emptiness must then exist either conventionally or ultimately.
The latter, as we have seen, is impossible, since then it would fail itself to be empty, and not only would a central tenet of Madhyamaka philosophy be contradicted, but the remainder would be rendered incoherent as well. But positing emptiness as a conventional existent and as the object of a correct view is no better. For things that appear conventionally appear as entities—as phenomena that exist independently and substantially. And all such appearance is, from the standpoint of Madhyamaka, in an important sense, false appearance.
Now, there is an important sense in which much of conventional reality also appears to us truly—that is, truly in conventional terms. So, for instance, this paper is in fact white, and so it appears. But note that the sense in which "This paper is white" is literally true is this: the perspective from which this paper exists is the conventional perspective, and in that perspective it is in fact white. But it is not from that perspective—the conventional perspective—that emptiness exists.
The perspective from which things are empty is not the conventional, but the ultimate perspective, and from that perspective, nothing exists, since nothing exists ultimately. To put this point another way, true predication is always predication from a perspective in which the subject of the predicate exists, and within which the predicate can be instantiated.
For conventional entities, the conventional standpoint provides such a perspective. But for emptiness, neither the conventional nor the ultimate standpoint can do the job: from the conventional standpoint there is no emptiness; from the ultimate standpoint there are no entities at all. Now let us consider the second entailment, concerning the manner in which emptiness would need to be viewed. Views are views of the things under description, and as having some nature. I view this paper as paper, as white, as a bearer of print, a product of a tree, and so forth.18
And again, so long as I am characterizing a conventional entity as it is viewed from the conventional perspective, there is no problem here. But when we attempt to extend this analysis to emptiness itself, problems again arise. For the attribution of properties, descriptions under which things can be viewed, again requires the existence of the substrata and the possibility of their serving as property bearers—as well as the dualism between substratum and property this presupposes.
The perspective from which this existence and this dualism are available is the conventional perspective, for only conventions bring ontology into play. But again, from that perspective, we find no emptiness: we find all kinds of entities, but we find them as entities. But from the perspective from which we find emptiness, we find no entities nor any characteristics, nor even emptiness itself or the fact of its emptiness. Hence again, since we cannot view emptiness even as empty, by virtue of its very emptiness, we cannot have a view of emptiness. This point is made pithily in a verse quoted by Nagarjuna in his autocommentary to the Vigrahavyavartam:19 "By their nature, things are not determinate entities. For they have only one nature, i.e., no nature" (Astasahasrikd Prajnaparamita).
CONVENTIONAL TRUTHS AND ULTIMATE TRUTHS
Up to this point, I have been relying heavily on the Madhyamaka distinction between the Two Truths, and on some standard Madhyamaka doctrine concerning each of them. It is now important in order to develop the positive characterization of our cognitive relation to emptiness and the coherence of positionlessness and the renunciation of all views to draw that distinction more explicitly. The distinction between the Two Truths is at its foundation a distinction between two kinds of objects of knowledge:
The conventional truth has as its domain all entities, properties, and so on posited by ordinary, unreflective awareness, science, philosophy, and other human activities. Its ontology is determined by language and by conceptual activity. The things we say about these phenomena, to the extent that they are true by the standards of human convention, are conventionally true. Ultimate truth, on the other hand, is truth independent of convention. Its domain is therefore the class of entities that exist independently of convention— of phenomena as they are independent of language or the ontology induced by language and thought. The fact that emptiness is what we find on such an analysis makes it legitimate to say that emptiness is an ultimate truth (indeed the only one) and the domain of ultimate truth.20'
Here it is necessary to pause a bit over the English terms "conventional entity/ truth/existent" and distinguish and comment on the Sanskrit and Tibetan terms these phrases translate.
There are two Sanskrit phrases translated as "convention." The most common is samvrti. This term enjoys one of the most delightful ambiguities of any philosophical term in any tradition. For it can literally mean "convention" in all of the senses that term has in English—"agreement," "coming together," "nominal," "ordinary," "everyday," and so on.
But it can also mean obscured, concealed, occluded, disguised. Plays on this ambiguity are significant in Madhyamaka literature as it is emphasized that conventional entities conceal their nature—the fact that they are merely conventional, and hence empty—and masquerade as substantial entities. The other term is vyavahara, which more accurately means "transactional," "linguistically determined." Unfortunately, though the latter term does not convey the ambiguity of the former, the former is often used in its "nominal" sense.
Tibetans also use two terms here, etymologically distinct from one another, reflecting the two meanings of samvrti, though often, because they are used to translate this term, regarded as synonymous. The first, tha snyad, really means "nominal," the second, kun rdzob, means "concealer"—literally "costumed." The former is almost always used to translate vyavahara, but both are used indifferently to translate samvrti. Tibetan Madhyamaka texts also use two distinct combinations of each of these terms as adjectival with noun phrases: each can modify either yod pa, "existent," or bden pa, "truth." So we can get phrases that mean "concealer," "truth," a "concealing existent," "nominal truth of nomimal existent."
When we say that something is a tha snyad yod pa, we are saying that it is something whose existence is entirely nominal, due to verbal conventions. We are saying of it that its status as an entity, as an enduring thing, as a referent of linguistic expression, derives entirely from the ontology induced by language and conceptual thought. Absent these conventions, we are saying; "Nature presents no joints at which to be carved, and a fortiori none by virtue of which this thing must be served as a portion to experience."
Nagarjuna argues that we can say this about anything. When we say of a thing that it is a kun rdzob yod pa, on the other hand, we are saying (i) that it is a something that appears to exist for a deluded consciousness (one that has not understood that in reality all phenomena are empty) or (2) that it presents itself falsely as existing inherently. Analysis shows that it does not exist in the manner that we instinctively posit it as existing. Nagarjuna would argue that all conventional entities have this character. These two terms hence appear to be coextensive.
There is, however, one point at which they diverge in application, and that is the point central for our discussion, that is, the status of emptiness itself. Emptiness is a tha snyad yod pa, but not a kun rdzob yod pa. First, why is it a tha snyad yod pa? Simply because emptiness is always the emptiness of something—of a table, of the parts of stages of a table, of the emptiness of that table, or if its emptiness. Independent of empty phenomena, there is no emptiness.23
The emptiness of emptiness just is its lacking inherent existence, and so existing nominally. But second, it cannot be a kun rdzob yod pa. Why not? Just because emptiness, when it is perceived by the kind of consciousness capable of perceiving it, as opposed to inferring its presence, does not come in costume. It does not deceive and is not apparent to a deceived consciousness.
When we see emptiness, we see what really is the nature of things; when we see things as empty, we shed delusion. It is in this sense that emptiness is not a conventional entity. But that does not contradict its status as conventional in the sense of being merely nominal. But being merely nominal is a way of not being an entity despite entering the field of knowledge. And this provides the key to a positive understanding of nonassertorial knowledge of the nature of things.
EMPTINESS AS A KNOWABLE II: EMPTINESS NEED NOT BE AN OBJECT OF KNOWLEDGE
We are approaching the conclusion—one poised on the edge of paradox and incoherence—that emptiness can be known but that it is not a possible object of 62 Madhyamaka knowledge, that we can say true things about it, but that those assertions in som sense indicate no proposition. To make sense of this conclusion, it will be necessary to remain conscious of the Two Truths and the two perspectives they provide from which emptiness may be known. Let us begin with some remarks of Candrakirti from his commentary on Nagarjuna's Mulamadhyamakakarika. He begins by considering the position that emptiness is an actual essence of all things:
Since entities are said to be empty, the essencelessness of entities does not exist. So, on the basis of their emptiness, entities are said to have a nature. This is explained to make no sense:
7 If there were even a trifle nonempty,
Emptiness itself would be but a trifle.
But not even a trifle is nonempty.
How could emptiness be an entity?
If emptiness existed, then essences of things would exist as its basis, but it doesn't. This is because emptiness and selflessness themselves are the universal characteristics of all things. It should be understood that because of the nonexistence of any non-empty phenomenon and of the non-existence of emptiness, since there is no non-empty phenomenon, it follows that emptiness and existence do not exist. It is dependent upon its opposite, and without that, like a sky-flower-garland, emptiness along with existence are explained not to exist. Since emptiness does not exist, entities and existence do not exist as its basis ... (833)
8 The victorious ones have said
That emptiness is the relinquishing of all views.
For whomever emptiness is a view,
That one will accomplish nothing.
So here emptiness is the ceasing of the perception of all views and of the persistence of all attachment to them. And since this is so, the mere relinquishing of views is not even an existent. We will not debate with whoever insists on seeing emptiness as an entity. Therefore, if one opposes this presentation through conceptual elaboration, how will liberation be achieved?
It is like this: Suppose a someone says to some one else "I have nothing" and he says in reply "I'll take that very nothing you say you have." If so, without anything there, by what means can we get him to understand that there is nothing there? In just this way, how can someone stop insisting on seeing emptiness as an entity through seeing emptiness as an entity? Therefore, the great doctors with great wisdom and realisation, who have performed great medical deeds, understanding their illness pass by them and refuse to treat them.
As it is said in the great Ratnakuta-sutra, "Things are not empty because of emptiness; to be a thing is to be empty. Things are not without defining characteristics through characteristiclessness; to be a thing is to be without a defining characteristic. Things are not without aspiration because of aspirationlessness; since to be a thing is to be without aspiration, whoever understands each things in this way, Kasyapa, will understand perfectly how everything has been explained to be in the middle path. Kasyapa whoever conceives of emptiness through objectifying it falls away from understanding it as it has been explained to be. Kasyapa, it would be better to view the self as just as stable as Mt Meru than to view emptiness in this way. This is because, Kasyapa, since emptiness is understood through the relinquishing of all views, whoever conceives of emptiness through a view, I have explained, will be incurable.
Kasyapa, consider this example: If a doctor gives a patient medicine, and this medicine cures all of his illness, but stays in his stomach, do you think that suffering will not arise, Kasyapa? Do you think this man will be relieved of the illness in his belly? No way, blessed one! If the medicine, having cured all of his illnesses, stays in his stomach, this man will certainly become seriously ill.
The Blessed one said, "Kasyapa, you should see the insistence on any view in just this way. If emptiness is seen like that, Kasyapa, whoever sees emptiness like that, will be incurable. I have said that it is like that. (83b-84a) Candraklrti makes two related points here. First, if emptiness were an entity, things would in fact have essences—namely the essence of being empty, which would contradict their very emptiness since that is denned as the absence of essence.24
Second, if emptiness were an entity, its relation to empty things would be contingent. It would then be possible for there to be non-empty things. But analysis that shows all things to be empty precludes that. But Candraklrti, like Nagarjuna, surely does not shy away from using the term emptiness or talking about emptiness itself. Conventionally, we can speak about anything, even emptiness. The question about such discourse that needs to be posed, however, is this: does the use of the term or the sense of that discourse implicate the existence of emptiness? The answer to this is a definite "yes and no." From a conventional perspective, emptiness is an existent. That is, if we ask, in ordinary philosophical discourse, whether things are empty, the answer is, for a madhyamika, "yes." Does this mean that they have the property of emptiness? Well, just as the conventionally true sentence that this paper is white entails that it has whiteness, this entailment seems perfectly good. This suggests that within the context of convention, emptiness is no more or no less real than other properties. They are conventionally existent and ultimately empty, and so is it. And this is exactly the observation that underlies madhyamika's emphasis on the nonduality of the conventional and the ultimate, and the insistence on the part of the dGe lugs interpreters on the existence of emptiness. But there is an important disanalogy that must not be overlooked: whiteness, like the paper, only appears at all from the conventional viewpoint. Seen from the ultimate standpoint, there is no paper, nor any whiteness. There is, from that standpoint, only their emptiness. So the fact that the paper is white is true from the very standpoint in the context of which the paper and its whiteness exist. To the extent that either is real, the assertion says correctly that the former is characterized by the latter. But all assertion—explicit in language or silent in thought—is nominal and conventional, including the assertions that the paper is empty and that emptiness exists. And from the conventional perspective, while emptiness can be known inferentially, it does not appear as it is. It is rather from the ultimate perspective that emptiness actually appears as emptiness. This is a delicate point, so let us take some case in articulating it. Paper and whiteness appear as, and are, paper and whiteness conventionally. So, to the extent that one cognizes either in a manner appropriate and adequate to such objects, one's knowledge is conventional. But emptiness can be known in two ways—indirectly through inference, that is conventionally; or directly, through that direct insight into the nature of reality that is the goal of Mahayana Buddhist practice. It is in the latter kind of gnosis that emptiness is known as emptiness, just as it is in conventional cognition that paper and whiteness are known as paper and as whiteness. But in this direct apprehension, there is no object, no entity to be perceived. For even emptiness is perceived by such a perceiver as empty, and hence not as an entity. Now to be sure, the emptiness that is perceived in these two ways is the same emptiness. In the one case, it is perceived as an existent quality; in the second as a mere negation—a mere absence of any essence or nature. But insofar as emptiness is the ultimate nature of things, it is the latter perception that gives us emptiness as it is. So there are certain respects in which emptiness is just like any other conventional existent—it can be the object of a conventional, conceptual consciousness and can be truly asserted to characterize conventional phenomena. Yet as such, it is merely nominal, dependently arisen, and hence empty of inherent existence. To say this is merely to say that, like paper and whiteness, it is a tha snyad yod pa. On the other hand, there is an important respect in which emptiness is wholly different: while conventional phenomena, even when apprehended correctly, are deceptive with regard to their ultimate nature, their emptiness, and hence are properly also characterized as kun rdzob yod pa, emptiness, when apprehended correctly, is not so deceptive. That is the sense in which it is an ultimate truth, even if not an ultimate existent. But for it to be nondeceptive in its appearance is for it to appear as a mere negation, and not as a positive object or an entity, and hence not as something that can serve as the subject of characterization—as the subject of proposition—or as the object of a view.
WHY ASSERTIONS OR REAL POSITIONS ABOUT EMPTINESS WOULD BE IMPOSSIBLE
Nonetheless, it must be possible to talk one's way into the ultimate perspective. The whole point of Madhyamaka philosophy is to do so. That is why Nagarjuna, for instance, writes:
10 Without a foundation in the conventional truth
The significance of the ultimate cannot be taught.
Without understanding the significance of the ultimate,
Liberation is not achieved. (MK 24)
Conventional truth, including the language and conceptual apparatus that mediates our experience and knowledge, the ontology they induce, and the objects and properties constituted within that ontology, form the ladder that allows us not only to comprehend the conventional world, but to comprehend its merely conventional character—that is, its ultimate emptiness. But that ladder must be kicked away. We cannot simultaneously remain within the conventional perspective and transcend it. But transcending it requires the transcendence of many things: we transcend the ontology of substance and attribute; the duality of subject and object; the commitment to the independent existence of objects of thought as well as that of thought. We enter into a vision of reality as an interdependent realm of essenceless relata. From such a perspective, insofar as the role of assertions is to predicate attributes of entities and to characterize objects of a subject's consciousness, language becomes curiously self-undermining. From that perspective, nothing can truly be said (at least not without contradiction).25 So, to say something about emptiness is necessarily to operate from the conventional perspective. And this is the perspective from which we must operate in order analytically to engender the cognition that allows such transcendence, but whose content, as Wittgenstein put it, can only be "shown" and "not said." That perspective contains rich descriptive and rhetorical resources. From it we can say true things about tables, elephants, and philosophical positions. And the true things we say can be useful not only in everyday transactions and philosophical debate, but also soteriologically: that is, they can be the very vehicles by means of which we show ourselves and others that which these assertions themselves assert the impossibility of ever asserting. But the things we say from that perspective can't be true of emptiness. And that is simply because from that perspective we get to say things about things, and about their properties, and emptiness simply is not one of those, though by treating it as though it were, we can move ourselves toward an inexpressible understanding of it.26
That is, to return to the language of positions and propositions: while we can correctly utter sentences that have "emptiness" in the subject (or predicate) position, for example, "Emptiness is empty," the analysis of the meaning of these sentences cannot involve the notion that an entity, emptiness is characterized by a property, emptiness. Hence, in the technical sense of proposition in the sense oipratijna, these sentences assert no proposition.27
Rather, they must be understood as pure denials—denials that phenomena, including emptiness, have any nature—showing the way that things are when seen per impossibile, sub specie aeternitatis.28
How TO UNDERSTAND POSITIONLESSNESS AS NON-SELF-REFUTING
In conclusion, I have emphasized the following features of Nagarjuna's account: (i) To say that the madhyamika is without position, asserts no proposition, has no view is neither to take a position, nor to assert a proposition, nor to have a view. That is, this is not a self-refuting position, because a position in the relevant sense is a position regarding the nature of things, a proposition is an extralinguistic entity that stands as the semantic value of a sentence, and a view is a view of an entity. And according to Nagarjuna and his followers, the function of language is not to characterize the nature of things, because things have no nature; language is not meaningful because of a correspondence to language-independent phenomena, because there are none; and an exposition of emptiness is not the characterization of an entity because it is the exposition of the incoherence of the very notion of an entity. (2) All of this is closely bound up with the question of whether emptiness is to be conceived of as an entity—as a positive phenomenon or an object of knowledge. I have argued that it cannot be coherently so conceived and that Nagarjuna never intends it to be so conceived. (3) This is not a mystical posit of an ineffable truth; it is a recognition of the inability to make assertions from a nonperspectival perspective, together with the recognition that perspectives are ontologically determinative. We can say that it is an anticipation of the doctrine of the hermeneutical predicament. Candrakirti expresses the point this way: So here emptiness is the ceasing of the perception of all views and of the persistence of all attachment to them. And since this is so, the mere relinquishing of views is not even an existent. We will not debate with whoever insists on seeing emptiness as an entity. Therefore, if one opposes this presentation through conceptual elaboration, how will liberation be achieved?
It is like this: Suppose a someone says to some one else "I have nothing" and he says in reply "I'll take that very nothing you say you have." If so, without anything there, by what means can we get him to understand that there is nothing there? In just this way, how can someone stop insisting on seeing emptiness is an entity through seeing emptiness as an entity? Therefore, the great doctors with great wisdom and realisation, who have performed great medical deeds, understanding their illness pass by them and refuse to treat them.
As it is said in the great Ratnakuta-sutra, "Things are not empty because of emptiness; to be a thing is to be empty. Things are not without defining characteristics through characteristiclessness; to be a thing is to be without a denning characteristic. Things are not without aspiration because of aspirationlessness; since to be a thing is to be without aspiration; whoever understands each things in this way, Kasyapa, will understand perfectly how everything has been explained to be in the middle path. Kasyapa whoever conceives of emptiness through objectifying it falls away from understanding it as it has been explained to be. Kasyapa, it would be better to view the self as just as stable as Mt Meru than to view emptiness in this way. This is because, Kasyapa, since emptiness is understood through the relinquishing of all views, whoever conceives of emptiness through a view, I have explained, will be incurable.
Kasyapa, consider this example: If a doctor gives a patient medicine, and this medicine cures all of his illness, but stays in his stomach, do you think that suffering will not arise, Kasyapa? Do you think this man will be relieved of the illness in his belly? No way, blessed one! If the medicine, having cured all of his illnesses stays in his stomach, this man will certainly become seriously ill.
The Blessed one said, "Kasyapa, you should see the insistence on any view in just this way. If emptiness is seen like that, Kasyapa, whoever sees emptiness like that, will be incurable. I have said that it is like that. (83b-84a) To say that things are natureless can only be designation—a ladder to be kicked away once we have ascended it. Indeed, if an assertion of positionlessness were meant to be the claim that one says nothing, that no words were uttered, or even that nothing was intended by those words, it would be self-refuting. But this is not what Nagarjuna intends—though nor does he mean merely that he has no false view, or that he ever says anything that is meant itself to be inherently existent. The claim to positionlesssness is deeply significant and represents a compelling vision of the utility of necessarily perspectival language and thought as well as of the limitations imposed by that perspectival character. That is, such assertions are meant as pointers—and that is the only way that such fundamental philosophical discourse can be understood—if we are to use it to get to, as Sellars calls it, an arche beyond discourse, the relinquishing of all views (Sellars 1997).
Source